Instructs the speech recognizer on how to process the audio content.
JSON representation |
---|
{ "audioEncoding": enum ( |
Fields | |
---|---|
audioEncoding |
Required. Audio encoding of the audio content to process. |
sampleRateHertz |
Required. Sample rate (in Hertz) of the audio content sent in the query. Refer to Cloud Speech API documentation for more details. |
languageCode |
Required. The language of the supplied audio. Dialogflow does not do translations. See Language Support for a list of the currently supported language codes. Note that queries in the same session do not necessarily need to specify the same language. |
enableWordInfo |
If |
phraseHints[] |
A list of strings containing words and phrases that the speech recognizer should recognize with higher likelihood. See the Cloud Speech documentation for more details. This field is deprecated. Please use |
speechContexts[] |
Context information to assist speech recognition. See the Cloud Speech documentation for more details. |
model |
Optional. Which Speech model to select for the given request. For more information, see Speech models. |
modelVariant |
Which variant of the |
singleUtterance |
If |
disableNoSpeechRecognizedEvent |
Only used in |
bargeInConfig |
Configuration of barge-in behavior during the streaming of input audio. |
enableAutomaticPunctuation |
Enable automatic punctuation option at the speech backend. |
optOutConformerModelMigration |
If |
AudioEncoding
Audio encoding of the audio content sent in the conversational query request. Refer to the Cloud Speech API documentation for more details.
Enums | |
---|---|
AUDIO_ENCODING_UNSPECIFIED |
Not specified. |
AUDIO_ENCODING_LINEAR_16 |
Uncompressed 16-bit signed little-endian samples (Linear PCM). |
AUDIO_ENCODING_FLAC |
FLAC (Free Lossless Audio Codec) is the recommended encoding because it is lossless (therefore recognition is not compromised) and requires only about half the bandwidth of LINEAR16 . FLAC stream encoding supports 16-bit and 24-bit samples, however, not all fields in STREAMINFO are supported. |
AUDIO_ENCODING_MULAW |
8-bit samples that compand 14-bit audio samples using G.711 PCMU/mu-law. |
AUDIO_ENCODING_AMR |
Adaptive Multi-Rate Narrowband codec. sampleRateHertz must be 8000. |
AUDIO_ENCODING_AMR_WB |
Adaptive Multi-Rate Wideband codec. sampleRateHertz must be 16000. |
AUDIO_ENCODING_OGG_OPUS |
Opus encoded audio frames in Ogg container (OggOpus). sampleRateHertz must be 16000. |
AUDIO_ENCODING_SPEEX_WITH_HEADER_BYTE |
Although the use of lossy encodings is not recommended, if a very low bitrate encoding is required, OGG_OPUS is highly preferred over Speex encoding. The Speex encoding supported by Dialogflow API has a header byte in each block, as in MIME type audio/x-speex-with-header-byte . It is a variant of the RTP Speex encoding defined in RFC 5574. The stream is a sequence of blocks, one block per RTP packet. Each block starts with a byte containing the length of the block, in bytes, followed by one or more frames of Speex data, padded to an integral number of bytes (octets) as specified in RFC 5574. In other words, each RTP header is replaced with a single byte containing the block length. Only Speex wideband is supported. sampleRateHertz must be 16000. |
SpeechContext
Hints for the speech recognizer to help with recognition in a specific conversation state.
JSON representation |
---|
{ "phrases": [ string ], "boost": number } |
Fields | |
---|---|
phrases[] |
Optional. A list of strings containing words and phrases that the speech recognizer should recognize with higher likelihood. This list can be used to:
See the Cloud Speech documentation for usage limits. |
boost |
Optional. Boost for this context compared to other contexts:
Dialogflow recommends that you use boosts in the range (0, 20] and that you find a value that fits your use case with binary search. |
BargeInConfig
Configuration of the barge-in behavior. Barge-in instructs the API to return a detected utterance at a proper time while the client is playing back the response audio from a previous request. When the client sees the utterance, it should stop the playback and immediately get ready for receiving the responses for the current request.
The barge-in handling requires the client to start streaming audio input as soon as it starts playing back the audio from the previous response. The playback is modeled into two phases:
No barge-in phase: which goes first and during which speech detection should not be carried out.
Barge-in phase: which follows the no barge-in phase and during which the API starts speech detection and may inform the client that an utterance has been detected. Note that no-speech event is not expected in this phase.
The client provides this configuration in terms of the durations of those two phases. The durations are measured in terms of the audio length from the start of the input audio.
The flow goes like below:
--> Time
without speech detection | utterance only | utterance or no-speech event
| |
+-------------+ | +------------+ | +---------------+
----------+ no barge-in +-|-+ barge-in +-|-+ normal period +-----------
+-------------+ | +------------+ | +---------------+
No-speech event is a response with END_OF_UTTERANCE without any transcript following up.
JSON representation |
---|
{ "noBargeInDuration": string, "totalDuration": string } |
Fields | |
---|---|
noBargeInDuration |
Duration that is not eligible for barge-in at the beginning of the input audio. A duration in seconds with up to nine fractional digits, ending with ' |
totalDuration |
Total duration for the playback at the beginning of the input audio. A duration in seconds with up to nine fractional digits, ending with ' |